A new term in MS research that has emerged in recent years is PIRA, or progression independent of relapse activity. It purports to quantify the proportion of disability worsening due to non-inflammatory neurodegenerative processes. However, the new metric has its shortcomings and the methodology used to quantify PIRA has varied since the concept first appeared in 2017.
At first glance, PIRA would appear to be similar to “smouldering MS’ or “silent progression”, although the three concepts are quite different. Smouldering inflammation or demyelination generally refers to chronic active and slowly expanding MRI lesions (Elliott et al. Brain 2019;142:2787-2799; free full text at www.ncbi.nlm.nih.gov/pmc/articles/PMC6736181/pdf/awz212.pdf). “Silent progression” describes the worsening disability that occurs independently of relapses or white-matter lesions and which appears to be correlated with brain atrophy (Cree et al. Ann Neurol 2019;85: 653–666).
In contrast, PIRA does not take into account MRI activity. The term originated in an analysis of data from the Tysabri Observational Program (TOP) of patients on natalizumab for a median of two years (Kappos et al. Mult Scler 2018;24:963-973). The study objective was to determine if using a roving rather than a fixed EDSS value would be more sensitive in detecting EDSS worsening. A fixed EDSS score (used in clinical trials) sets the benchmark at baseline. A roving EDSS score resets the baseline to the current value so that the comparison changes over time. For example, if a baseline EDSS score of 2.5 improved to 1.5 before returning to 2.5, there would be no net change from baseline and no disability worsening would be recorded. With a roving EDSS, the score would be rebaselined as 1.5 and the follow-up score of 2.5 would be recorded as a disability worsening event. While a roving EDSS may be more sensitive to disability worsening, it arguably overcounts periodic fluctuations in disability scores as disability worsening. Thus, in the TOP analysis, the proportion of natalizumab-treated patients with 24-week confirmed disability worsening (CDW) was 20.3% with a fixed EDSS baseline, and 37.1% with a roving EDSS value. The proportion with PIRA was 10.1% and 24.1%, respectively, over 288 weeks. This represents about 50% and 65% of all CDW events. An unexplained finding is that PIRA as a proportion of all CDW events actually decreased over time, whereas the opposite would be expected.
A roving EDSS was also employed in an analysis of the OPERA studies of ocrelizumab versus interferon-β-1a (Kappos et al. ECTRIMS 2017; abstract P654). For added complexity, a composite disability measure was used to assess 24-week confirmed worsening on EDSS, Timed 25-Foot Walk (T25FW) or 9-hole Peg Test (9HPT). To be labelled as PIRA, there could be no relapse in the 30 days before or after the disability worsening. Since disability was evaluated every three months, there was a one-month window between disability assessments during which a relapse or a change in T25FW/9HPT could occur without disqualifying PIRA.
The proportion of patients with 12-week confirmed composite worsening was 21.5% with ocrelizumab versus 30.7% with IFN-β-1a. The 12-week CDW rate (EDSS only) was 9.8% versus 15.2%, respectively, which was closer to the CDW rates (pooled 9.1% vs. 13.6%) reported in the clinical trial (Hauser et al. N Engl J Med 2017;376:221-234). The composite results confirmed at 24 weeks were 16.1% vs. 22.6%. For EDSS only, results were 7.6% vs. 12.0% (vs. 6.9% and 10.5% in the trial with a fixed EDSS). Interestingly, use of a roving rather than a fixed EDSS had a greater negative impact on CDW results for IFN-β-1a compared to ocrelizumab.
The two analyses concluded that with natalizumab, about 65% of disability worsening was not associated with relapses (PIRA). With the composite EDSS, PIRA represented about 85% of CDW events with IFN-β-1a and 92% with ocrelizumab. Using the more conservative assessment of EDSS (rather than the composite measure), about 75% of CDW events could be attributed to PIRA.
Two other groups have also conducted their own PIRA analyses. The Swiss MS Cohort study (n=917) used a roving EDSS over a median 4.6 years of follow-up (Lorscheider et al. ECTRIMS 2019; abstract 273). Overall, 198 patients (22%) experienced CDW, including 131 (14%) who met PIRA criteria over a 6-year period; these results were comparable to those reported in the TOP analysis (20.3% CDW, 10.1% with PIRA-EDSS over a similar time period). Among those with CDW, 62% of the events were attributed to PIRA in the CIS/MS group. In the SPMS/PPMS group, 93-95% of CDW events were attributed to PIRA, which may be due to the lower benchmark needed (0.5-point EDSS change) and the fact that PIRA essentially defines progressive MS. As expected, the overall number of CDW events/100 patient-years was lower among patients treated with a MAb (3.3/100 PY) compared to an oral (4.5) or an injectable (4.8) DMT or who were untreated (6.0). It is interesting to note that the number of PIRA events/100 PY was similar in patients treated with a MAb (2.2) or oral (2.5) DMT (vs. 3.1 for injectables and 3.3 for untreated). Between-group differences were not significant, which likely reflects the low absolute number of CDW or PIRA events over the 6-year follow-up. It is unclear how the frequency of dosing and the timing of EDSS assessments influence PIRA results.
A second Swiss database analysis compared fingolimod with injectable DMTs and used more rigorous PIRA criteria: 1-year confirmed EDSS worsening and no relapses during the 3-4-year follow-up (von Wyl et al. ECTRIMS 2019; abstract P751). Overall, only 6.6% of this cohort remained relapse-free. CDW was reported in 7.6% in the fingolimod group versus 8.8% in the injectable group. The overall rate of PIRA was 4.1% with fingolimod and 3.1% with injectables. As in the previous study, a higher-efficacy DMT did not provide additional benefit with respect to PIRA. The caveat is that >90% over the observation period experienced a relapse so the proportion with PIRA was necessarily low.
These studies show that PIRA can be variously defined as worsening disability (with EDSS or a composite) independent of relapses (within a defined period or in relapse-free patients). The key finding appears to be that a high proportion of CDW events occur in the absence of relapses, although the number of PIRA events in absolute terms is low in CIS/RRMS (perhaps 5%/year). This may represent the stratum of neurodegeneration that is largely unaffected by DMTs that target CNS inflammation.
PIRA is an evolving concept and, like other composite metrics (such as NEDA), it requires a standardizing of its definition and methodology. It implies that any disability that is associated with inflammation is not “true progression”, which may not be correct. Perhaps its greatest shortcoming is not including MRI findings. To quantify the underlying component of progression resulting from chronic neurodegeneration, which may be masked by short-term disability worsening due to inflammation, it may be useful to examine progression independent of MRI activity (PIMA? PIRA-2?).